Operant reinforcement versus reward expectancy: effects on anticipatory eye movements

Previous studies have shown that smooth pursuit can be modulated by reward either during transient target blanking (Madelain and Krauzlis, 2003), or during pursuit of an ambiguous stimulus (Sch├╝tz et al, 2015). Moreover, anticipatory smooth pursuit eye movements (aSPEM) are observed before target appearance when predictive information about target motion is available, reducing the typical sensorimotor delay between target motion onset and foveation. By manipulating the probability of target motion direction we previously biased the direction and mean velocity of aSPEM, suggesting that motion-direction expectancy has a strong effect on the initiation of aSPEM. To further understand the nature of anticipatory smooth eye movements, we now investigate different effects of reinforcement on aSPEM with two distinct experiments. In a first experiment, we manipulated the proportion of rewarded trials (signaled by a green-color target associated to a monetary gain) across motion directions (right versus left) in separate experimental blocks maintaining the total number of rightward and leftward trials equal. Our results indicate that the proportion of rewarded trials does not significantly affect anticipatory eye movements. In a second operant conditioning paradigm, the reward was contingent on anticipatory pursuit: in separate blocks with fixed probabilities of rightward (vs leftward) trials, monetary reward was contingent on a criterion-matching anticipatory velocity measured before target motion onset. Contingent monetary reward increased (or decreased) anticipatory velocity compared to the probability-biased condition. To further probe the effects of operant conditioning, we used a yoked-control procedure in a second group of subjects, in which the relation between pursuit and reward was randomized. We observed a reduction of aSPEM mean velocity. Overall, these results strengthen the notion that eye movements (including anticipatory ones) can be considered as an operant behavior, whereas the expectancy for a non-contingent reward cannot efficiently biased them.

Meeting abstract presented at VSS 2016


All material (c) L. Perrinet. Please check the copyright notice.

This work was supported by ANR project ANR-13-APPR-0008 "ANR R.E.M.".
ANR logo

TagYear16 TagPublicationsProceedings TagAnrRem

welcome: please sign in